ABSTRACT The major psychoses (SZ, SAD, BDP), when defined by clinical phenomenology alone, overlap extensively on neurobiological, biomarker, co-morbid, symptomatic, and genetic characteristics. Our field may benefit from transformational re-conceptualizations of disease seen in other areas of medicine when biological variables are considered in disease definitions and identification. This approach in psychiatry will depend on: (i) use of well- defined disease domains, (ii) large samples that capture clinical heterogeneity and support statistical approaches, and (iii) ability to acquire quantifiable laboratory measures to inform re-conceptualization of disease characteristics. The 5-site B-SNIP focus is psychosis, an ideal clinical phenotype for this purpose. B- SNIP1 recruited over 2500 volunteers and performed dense phenotyping across multiple levels of analysis (cognitive, psychophysiological, brain imaging, social and clinical). The overall data described a continuum of phenotypic alterations across the DSM psychosis diagnoses (BDP, SAD, SZ) with little evidence of diagnostic specificity. In an attempt to use these dense phenotypic characteristics to define biologically based subgroups, we re-grouped probands using biomarkers and a multistage multivariate analysis procedure. We identified 3 psychosis ?Biotypes? based on core phenotypic features. Biotypes showed unique differences across external validators that were not used in the initial construction of the categories. B-SNIP2 will replicate and extend B- SNIP1 using enhanced proband number, biomarker panel, and sophistication of multivariate statistical approaches. We will accomplish our goals within the context of two specific aims. SA(1) Construct a `Psychosis Biomarker Database' (PBD): Recruit 3000 new psychosis probands and 600 healthy volunteers and collect data including clinical, psychosocial, electrophysiological, ocular motor, imaging and blood biomarkers. Core biomarkers (used for Biotype definition) and external validators (used for verifying neurobiological distinctiveness of Biotypes) will be collected as specified. Genetic characteristics of the participants will be obtained in collaboration with the Broad Institute. SA(2) Contrast and test taxometric approaches to categorizing psychosis: Evaluate the ability of different toxonomic structures to define psychosis subgroups, based on data in the PBD: (i) DSM, (ii) B-SNIP2 biotypes based on clinical variables, (iii) B-SNIP1 Biotypes, (iv) B-SNIP2-generated biotypes based on biomarkers, and (v) B-SNIP2 biotypes based on both clinical variables and biomarkers. Beginning with traditional DSM diagnostic criteria as the taxonomy and testing (i)-(v) we will use linear, quadratic and nonparametric discriminant function analysis applied to external biomarker validators to examine the association between the traditional diagnostic system and the biologically- derived classification (imaging, psychosocial and genetic external validators). We will be able to determine the strongest taxonomic approach based on biological characteristics. We seek a rational classification of psychotic disorders that will be successful in identifying novel disease targets and treatments approaches.